Squeak-dev July 2002

squeak-dev@lists.squeakfoundation.org

205 participants
393 discussions

Reminder - ESUG Smalltalk Joint Event
by Noury Bouraqadi 29 Jul '02

29 Jul '02

Appologize for multiple receptions. -------------------------- The European Smalltalk User Group (www.esug.org) invites you to the Smalltalk Joint Event 10th ESUG Smalltalk Conference & 2nd Smalltalk Doctoral Symposium Tuesday, August 27, - Friday, August 30, 2002 & Camp Smalltalk Saturday, August 24 - Monday, August 26, 2002 Douai, France http://www.esug.org/summerschools/2002_Douai/index.htm (the final program and the registration process will be ready for the end of June) GENERAL INFORMATION For the 10th consecutive year, the European Smalltalk User Group is organizing a Smalltalk Conference to bring together Smalltalk users from both academia and industry in a friendly atmosphere. As always, the attendents will enjoy tutorials, workshops and demonstrations on various Smalltalk related subjects. Continuing a strong tradition in organizing Smalltalk events for Smalltalk developers, this year promises to be very exciting. All these exciting events will take place in the last week of August, at the Ecole des Mines de Douai in Douai, France (1h30 away from Brussels by car and 1h10 away from Paris by train). The Conference takes place at Douai in France. Early information can be found at www.esug.org or at http://www.iutc3.unicaen.fr:8000/esug2002/1. ESUG organizes a Smalltalk Joint Event comprised of: Camp Smalltalk, Smalltalk Conference, a Free Software Day, a Teachers Forum and Smalltalk Doctoral Symposium! - An official (and free) Camp Smalltalk (notice that it starts on the saturday before so that Campers can be productive and stay during the week). ESUG will sponsor the food. - The Smalltalk Doctoral Symposium, where students doing Master and PhD theses have the opportunity to present and discuss their research work with industrial and academic Smalltalk experts. - Finally the Advanced Seminars: Similar to last year, we will offer two tracks: invited talks which should interest everyone, and the technical presentations, which are more specicalized. Dave Thomas (the OTI founder) and Dave Simmons (SmallScript creator) are the keynotes speakers This year: - a free Software day will be organized around GnuSmalltalk, PocketSmalltalk, and Squeak (contact Serge.Stinckwich(a)info.unicaen.fr). - a Teachers Forum will be organized to share teaching experiences. We are looking for experience reports. We are looking for teachers to report their experiences using Smalltalk (contact ducasse(a)iam.unibe.ch) - ESUG will support the Squeak Foundation by donating 10% of its benefits. Exhibitors or sponsors should directly contact Stephane Ducasse (ducasse(a)iam.unibe.ch) or Roel Wuyts (wuyts(a)iam.unibe.ch) for more information. The conference fees has been lower to their minimum: 300 Eur for companies and 150 ? for academics. The hosting will be excellent with internet connection in the rooms: accomodation: 50 or 60 ? per night (site on progress: http://www.esug.org/summerschools/2002_Douai/registration.htm) IMPORTANT DATES Early registration: 15 of July (Please register yearly so that we can prepare ESUG) The organizing commitee -- ------------------------------------------ Noury Bouraqadi - Enseignant/Chercheur Ecole des Mines de Douai - Dept. G.I.P http://csl.ensm-douai.fr/noury ------------------------------------------

1 0

Re: [BUG] Preferences hanging on to obsolete classes
by Richard A. O'Keefe 29 Jul '02

29 Jul '02

I'm continuing this thread in the Squeak mailing list to bring home to people just how confusing things can be in the X world (and how really really nice it is to have a Squeak I can trust...) Ian Piumarta wrote helpfully: Unless there is some compelling reason not to, use xterm instead (AFAIK it's bundled with openwin -- if you haven't got one, I'll mail one to you). Then either pass "-leftbar" when starting xterm or put xterm*rightScrollBar: false in your `.Xresources'. The only reason I use dtterm is that that's what I got by default. However, the /usr/openwin/bin/xterm that comes with Solaris 2.8 does not accept a -leftbar command line option, nor does 'man xterm' acknowledge the existence of a rightScrollBar resource entry. The only version information I can find is that the binary is dated December 16, 1999. (rxvt may well support the same options, if you really despise xterm.) Despise xterm? I don't even understand it. There doesn't seem to be any rxvt here.

1 0

Re: A problem hindering SmallTalk's popularity:
by Richard A. O'Keefe 29 Jul '02

29 Jul '02

"i r thomas" <israelrt(a)optushome.com.au> wrote: Richard, I have the deepest respect for you (I used to hunt out and read your messages in comp.lang.prolog back in the 80's.) (I am assuming that you are the Richard O'Keefe of Quintus Prolog fame and who wrote The Craft of Prolog.) That's me. However, I am afraid that this is a trifle disingenuous. Although you are a novice at Morphic, you have far more computer science and software development experience than most people around. This does not diminish the feeling of lost helplessness that Morphic usually inspires in me. It certainly doesn't diminish my own fervent wish for something recognisable as adequate documentation in Squeak. (Of all the things that Quintus did, I think I am proudest of our documentation. There _still_ isn't anything to match it, even though our technology is now superseded.) I also want to stress that my computer science and software development experience does NOT extend to GUIs. The last GUI I really understood was the Xerox Interlisp-D system. Two chapters of the manual, that was all, and not very big ones. But *boy* could you get things done easily with those two chapters. One reason why I used to be distressed at the idea of Squeak dropping MVC was that I had put a lot of effort into trying to understand MVC and still wasn't there, and didn't want to waste that. So 'building a radio button' is *way* outside any expertise I may have. I've never even *used* a radio button before in any language, let alone built one. What did I have that was actually relevant to this problem? 1. I have read the Morphic tutorials. (Which DO exist. Big point.) 2. I have a basic familiarity with the Squeak IDE. (Tutorials also exist.) 3. I have a willingness to keep pushing on the grounds that I expect Squeak to have _something_ I can use. That's it, really. I would suggest that Lilly ( the student who had difficulty with radio buttons ) and her teacher are extremely unlikely to be within several orders of magnitude of your competence and knowledge. Lilly, yes, of course. Lilly's TEACHER doesn't get off that lightly. When I got my 4th year students onto Squeak, I hunted down a bunch of tutorials and stuck them in a directory and told them where to find them. (Hey, I got them for *me*, but they didn't need to know that!) As the Squeak books have come out, I have ordered them for our library. As the free books appear on Stephane Ducasse's web site, I slurp them down and put them in the same place as the tutorials, so that the students have access to them. These are activities that do not require any great competence or knowledge, but can be a big help to students. I spent a couple of hours this last weekend trawling through some of the Morphic sources. (Again. This time I started to understand a bit.) And there is MUCH that displeases me. For example, the morphic chapter I was going to work through starts with HeadingMorph. Object Morph Has class comment, of very little use. BorderedMorph NO CLASS COMMENT. RectangleMorph NO CLASS COMMENT. HeadingMorph NO CLASS COMMENT. MorphExample NO CLASS COMMENT. We've been told for years that Morphic was going to be rewritten. It hasn't happened. Lots of these classes have been around for quite a while and they still are not commented. So Lilly and co are RIGHT that the internal documentation desperately needs a lot of work. Some serious effort put in near the root of the 'Morph' hierarchy might be a good place to start. Or how about "everything in Morphic-Widgets", like BorderedStringMorph (no class comment), ColorPickerMorph (ditto), DropDownChoiceMorph (ditto), HaloSpec (ditto), JoystickMorph (ditto), NameStringInHalo (ditto), NewHandleMorph (ditto), PopUpChoiceMorph (ditto), ProgressMorph (ditto), RulerMorph (ditto), ScaleMorph (class comment exists but doesn't say what they are or are for), SimpleButtonDelayedMorph (no class comment), SimpleButtonMorph (ditto), SimpleSliderMoprh (ditto), SimpleSwitchMorph (ditto), StarMoprh (ditto), StringButtonMorph (ditto), TestBalloonFont (ditto), ThumbNailMorph (ditto), TickIndicatorMorph (ditto), TinyPaint (ditto), ToggleButtonMorph (ditto), TrashCanMorph (ditto), UpdatingStringMorphWithArgument (ditto), UpdatingThreePhaseButtonMorph (ditto), ZoomMorph (ditto). That is, nearly EVERYTHING that someone looking for useful widgets in Morphic might look at has NO CLASS COMMENT. This just is not anywhere near good enough. (Checked in 3.2.) However, the morphic chapter and tutorials DO exist and make up for a heck of a lot. Indeed, after playing around on the weekend, I could see more ways to make radio buttons in Squeak than I'll ever live long enough to try. Basically, ANY morph can be a button if you want it to be. If I can get my head around some of these morphs I might provide a few class comments myself. (Hey, if someone who _isn't_ an experienced Morphic programmer can even _consider_ providing useful class comments, that says something nearly as good about morphic as the need to do so says something bad.)

1 0

RE: Squeak and SQlite Database
by Jarvis, Robert P. (Contingent) 29 Jul '02

29 Jul '02

For a database used with Squeak (or any Smalltalk) I'd say that the important thing is that it should store objects, i.e. be an object database. Relational databases are too concerned with bits-and-bytes. Minnestore is a possibility, but its development has apparently ceased for the time being. Are there any other open-source, publicly available object databases for Smalltalk? Bob Jarvis Compuware @ Timken > -----Original Message----- > From: Stephen Pair [mailto:spair@acm.org] > Sent: Thursday, July 25, 2002 9:20 AM > To: squeak-dev(a)lists.squeakfoundation.org > Subject: RE: Squeak and SQlite Database > > > For me, the important thing for Squeak (as far as DBs are > concerned) is > to have structured (Btree, Hash, etc) access to secondary > storage. It's > also good to have some options regarding where the physical storage > resides and where the DB processing occurs (i.e. local or a remote > machine and disk) as well as transactions, logging, rollback, > etc. SQL > access is far less important (IMHO). > > So far, the only database that I've come across that meets these > fundamentals is BerkeleyDB...does anyone know of others that > are similar > in nature to BerkeleyDB (I'd like to have some options for what I'm > working on). > > In regard to a "platform independent common Squeak database," I think > that it should be qualified as a "platform independent common > Squeak SQL > database API"...meaning that rather than try and pick a single SQL > database for all platforms and applications, you specify an > API that is > sufficiently robust to handle a wide variety of SQL database > implementations. Also, we should be precise and call it an "SQL" > database API to distinquish it from the many other types of databases > that exist and would need their own API. > > - Stephen > > > -----Original Message----- > > From: squeak-dev-admin(a)lists.squeakfoundation.org > > [mailto:squeak-dev-admin@lists.squeakfoundation.org] On > > Behalf Of Torsten.Bergmann(a)phaidros.com > > Sent: Thursday, July 25, 2002 3:03 AM > > To: squeak-dev(a)lists.squeakfoundation.org > > Subject: Squeak and SQlite Database > > > > > > Is anybody using or planning to use the SQLite > > database with Squeak ? (http://www.hwaci.com/sw/sqlite/) > > > > From the page: > > > > SQlite is a C database engine in a small C library. > > Programs that link with the SQLite library can have SQL database > > access without running a separate RDBMS process. > > > > Features: > > - Implements a large subset of SQL92. > > - A complete database (with multiple tables and indices) is > stored in > > a single disk file. > > - Atomic commit and rollback protect data integrity. > > - Small memory footprint: less than 20K lines of C code. > > - Four times faster than PostgreSQL > > - very simple C/C++ Interface requieres the use of only three > > functions > > and an opaque structure > > - Build and tested under Linux and Win2000 > > - Source are uncopyrighted. Use for any purpose. > > > > Could this be the base for platform independent common squeak > > database ? > > > > Bye > > Torsten > > > > > > ********************************************************************** This message and any attachments are intended for the individual or entity named above. If you are not the intended recipient, please do not forward, copy, print, use or disclose this communication to others; also please notify the sender by replying to this message, and then delete it from your system. The Timken Company **********************************************************************

11 17

[BUG][UNIX][VM] Cursor>>beCursor (prim 101) fails
by Ned Konz 28 Jul '02

28 Jul '02

Using Ian's 3.2-5devel VM, in an MVC project, every time the cursor shape tries to change I get a walkback from a primitive failure. Cursor(Object)>>error: Cursor(Object)>>primitiveFailed Cursor>>beCursor Cursor class>>currentCursor: EventSensor(InputSensor)>>currentCursor: Cursor>>show Cursor>>showWhile: StandardSystemController>>adjustWindowBorders StandardSystemController>>checkForReframe StandardSystemController>>controlActivity StandardSystemController(Controller)>>controlLoop StandardSystemController(Controller)>>startUp [] in ControlManager>>activeController: [] in BlockContext>>newProcess -- Ned Konz http://bike-nomad.com GPG key ID: BEEA7EFE

2 3

Key events and special keys
by Luciano Notarfrancesco 28 Jul '02

28 Jul '02

I noticed that Morph|handleKeyDown: and Morph|handleKeyUp: don't get called when I press Shift, Alt Ctrl, the function keys, etc. Is there any way to handle those events? Does Squeak receive those events at all? Luciano.-

2 1

Re: [ANN] [UPDATES] 20 updates for 3.3a
by Bruce ONeel 28 Jul '02

28 Jul '02

Hi, Squeak3.3a-4934.zip is now at ftp://st.cs.uiuc.edu/SmalltalkSqueak/3.3alpha. cheers bruce Scott Wallace <scott.wallace(a)squeakland.org> wrote: > Here are the preambles for twenty updates for Squeak 3.3a that have > recently been released to the external 3.3a update stream. > > Cheers, > > -- Scott > > ---------------------- > > 4915resourceColorFix-raa -- Bob Arning -- 23 June 2002 > Republished 7/7/02 sw to the 3.3a update stream, with preamble updated. > When loading resource Forms, do NOT discard color table information. > It was only with the advent of update 4859 (3.3a) that this became an > issue." > > 4916dupKeys-nk -- Ned Konz -- 11 July 2002 > A persuasive argument was made by David Farber to allow duplication > of Ctrl and Alt keys. > This adds a preference called 'duplicateControlAndAltKeys' that does > just that. You can still just swap the keys if you want. > Per Doug Way's suggestion, this also changes the default keys to swap to be: > $c . $x . $v . $a . $s . $f . $g . $z > x (cut) > c (copy) > v (paste) > a (select all) > s (save) > f (find) > g (find again) > z (undo) > And (unlike the original change set) this will maintain existing > Preference settings when installed. > The two preferences are mutually exclusive; setting one will reset the other. > One small manual correction made in the Postscript by Scott Wallace 7/11/02. > > 4917dupNavBar-sw -- Scott Wallace -- 11 July 2002 > Removes the old navigator bar on reloaded projects if the current > image requests suppression of the navbar." > > 4918loadEToys-sw -- Scott Wallace -- 11 July 2002 > Removes barriers to loading etoys into 3.3. Caution: this update is > for 3.3 only." > > 4919reportFullStack-sw -- Scott Wallace -- 19 July 2002 > In the error report, print out the short form of *all* stack frames > in the same 'rcvr(class) >> selector' format, after printing out the > long form of the first four stack frames. This makes the entire > compact summary available in a single, readable tableau. Formerly, > the first four stack frames did not appear in the compact summary." > > 4920flapTuning-sw -- Scott Wallace -- 19 July 2002 > * Gives explicit object-names to the etoy flaps (Supplies & > Navigator) so that they show up in the morph halo that way, per > request from Kim Rose. > * Makes the Navigator shrink-wrapped, as per prescription from Andreas." > > 4921scriptParameters-sw -- Scott Wallace -- 20 July 2002 > Allows any user-defined etoy script to have a parameter. > To add a parameter, invoke 'add parameter' from Scriptor or Viewer menu. > Change parameter *type* from the Scriptor header, using up/down > arrows or clicking for menu. > Obtain a parameter tile by dragging from the parameter position in > the Scriptor header. > Caution: This applies to 'classic tiles' only. 'Universal tiles' are > not handled by this work. > Note: the following is a hand-jimmied class-definition for > ParameterTile to make it ok both on 3.2 and 3.3" > > 4922paramFixups-sw -- Scott Wallace -- 22 July 2002 > * Fixes a glitch in backing out of renaming a script -- we had not > been allowing the user to retain an existing name if it had a colon > in it. > * If you change a parameter type in such a manner that it introduces > a type error for any tile phrases in the script that called for the > parameter when it was a different type, the tiles fix themselves up > to keep everything type-rosy. > * Adds a menu item to TilePad's halo menu allowing you to restore > the default literal tile for the particular type. This allows you, > albeit somewhat obscurely, to revert a Parameter tile to a plain > numeric tile, for example." > > 4923optionalLexiconPane-sw -- Scott Wallace -- 23 July 2002 > Makes the custom Lexicon control buttons occupy a separate pane *in > addition to* the one governed by the optionalButtons preference. > Thus, if #optionalButtons is true, a Lexicon will have *two* rows of > control buttons -- one offering the custom lexicon-specific > navigation features and the other with the standard repertory of > senders, implementors, etc, which had for space reasons previously > had to be omitted from the Lexicon's touchable controls. > Also, fixes a seemingly-long-standing bug that appears to have made > protocol browsers not work properly in mvc." > > 4924findA-sw -- Scott Wallace -- 23 July 2002 > Consolidates the main work of six methods that service the various > windows-menu 'find-a' requests (e.g. 'find a Transcript') into a > single place. > One of these six is the newly added 'find a fileList', with cmd-L as > its desktop-cmd-key shortcut." > > 4925fixAllCat-sw -- Scott Wallace -- 24 July 2002 > Fix up methods that had '-- all --' as their home category" > > 4926versionBrowser-BG -- Boris Gaertner, Scott Wallace -- 25 July 2002 > Fixes a glitch that caused the creation of a spurious concrete method > category named '-- all --' when a method definition was reverted in a > versions browser that had been originally launched from a browser > that had had the '-- all --' category selected. > Note: Boris penetrated to the core of this bug, and then produced a > fix on a Squeak 2.7 system; Scott integrated Boris's fix with the > post-2.7 changes to the method, and did some further trimming." > > 4927condenseSources-sw -- Scott Wallace -- 25 July 2002 > Because the presence of spurious concrete '-- all -- method > categories breaks the condenseSources process, we now make a > preemptive call to the method that fixes up all such errors before > proceeding with the real work of condenseSources. Thanks to Boris > Gaertner for finding this problem." > > 4928chgSetForVersion-dew -- Doug Way -- 7 September 2001 > Adds a 'find original change set' menu item for the Versions Browser, > which locates the single changeset (or sources file) which originally > contained the selected method version. It does this by matching the > file pointers stored in the ChangeRecord and ChangeSet." > > 4929parameterMenuIcon-sw -- Scott Wallace -- 27 July 2002 > Use a black triangle as the type-menu icon on a parameter tile, as > per suggestion from Alan." > > 4930fasterSlideToTrash-sw -- Scott Wallace -- 26 July 2002 > When sliding large morphs to the trash, animate a thumbnail rather > than the full-sized morph. This speeds things up considerably." > > 4931versBrowsFix-sw -- Scott Wallace -- 27 July 2002 > Fixes a glitch in update 4926 that only showed up when browsing > versions of a method on the class side." > > 4932changeEnh-nk -- Ned Konz -- 26 June 2002 > This adds an item to the ChangeSorter category menu that lets you > make a change-set category with all the change sets that change a > particular class." > > 4933changedOthersMethods-nk -- Ned Konz <ned(a)bike-nomad.com> -- 30 March 2002 > This adds a choice to the shifted change set menu in ChangeSorters > and ChangeSetBrowsers (called 'check for any other authors') that > will find any methods with historical versions that don't have your > initials. > This can be used to find other people's code that you've modified." > > 4934chgSorterMisc-sw -- Scott Wallace -- 27 July 2002 > * Adds check-for-uncommented-classes to the checks section of the > shifted change-sorter menu. > * Adds cmd-hypen as kbd shortcut for subtract-other-side in the dual > change sorter. > * Adds ClassListBrowser, a browser variant that shows the code for an > arbitrary list of classes. See class comment or browse class-side > code for examples"

1 0

Re: [GOODIE] RadioButtonGroupMorph-nk
by Richard A. O'Keefe 28 Jul '02

28 Jul '02

Ned Konz provided a GOODIE from a 3.2gamma system. I'm running 3.2 final, latest update: #4917, and the code doesn't work because there is no UpdatingThreePhaseButtonMorphEnhanced class. Making #initializeWithValues: refer to ThreePhaseButtonMorph instead gets a little further, but not much, because the buttons you get from the class Squeak 3.2 _does_ have don't respond to #getSelector: Making it refere to UpdatingThreePhaseButtonMorph instead fixes that but leaves me staring at #getArgument: which is still not understood. Sorry, stuck.

2 1

New Random Number Generator
by Mark4Flies＠aol.com 28 Jul '02

28 Jul '02

Years ago I requested that the original random number generator inherited from Smalltalk-80 be updated to a modern linear congruential method. In no time at all, it was done! This time I thought I could help with a newer generator. I use Squeak intermittently so when I wanted to bring up a new simulation and realized that drawing a lot of random numbers would exhaust the current generator, I saw an opportunity to pick Squeak up again AND finally contribute something. The method is known as the Mersenne Twister and we use it in our commercial software at SAS Institute. You can find everything about it at the source: http://www.math.keio.ac.jp/~matumoto/emt.html I got the class to work (no syntax or logic errors) but the numbers are any thing but uniformly distributed on [0,1) ! I think the problem is related to how the algorithm coded in C relies on bit operations to work within unsigned 32-bit integers. I tried to constrain the elements of a WordArray (for example, masking all intermediate results with bitAnd: 16rFFFFFFFF) but so far it still is not cooking up good numbers. I was hoping that someone might have some experience with operations on arrays like this in C or in Squeak and they could give me a clue. In the mean time I will keep experimenting to learn more about the methods in Squeak. (I love this stuff!) Here is my new class, followed by the original C program. (The C program can be set up for random 32-bit integers or floating point numbers on [0,1), easily. Just look for two sets of comments related to this behavior. Right now it is set up for unsigned long integers.) BIG thanks in advance for any help that you might provide. -Mark Bailey 'From Squeak3.0 of 4 February 2001 [latest update: #3545] on 27 July 2002 at 11:27:51 am'! Random subclass: #MTRandom instanceVariableNames: 'mt mti ' classVariableNames: 'LowerMask M MatrixA N TemperingMaskB TemperingMaskC UpperMask ' poolDictionaries: '' category: 'Kernel-Numbers'! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/27/2002 11:00'! mask: aWord1 and: aWord2 "Perform bit logic and maintain unsigned four-word 32-bit integers" | t1 t2 | t1 _ (aWord1 bitAnd: UpperMask) bitAnd: 16rFFFFFFFF. t2 _ (aWord2 bitAnd: LowerMask) bitAnd: 16rFFFFFFFF. ^ (t1 bitOr: t2) bitAnd: 16rFFFFFFFF! ! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/27/2002 11:25'! temperingShiftL: anInteger "Second of four Mersenne Twister shifts" ^ anInteger >> 18 bitAnd: 16rFFFFFFFF! ! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/27/2002 11:25'! temperingShiftS: anInteger "Second of four Mersenne Twister shifts" ^ anInteger << 7 bitAnd: 16rFFFFFFFF! ! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/27/2002 11:25'! temperingShiftT: anInteger "Second of four Mersenne Twister shifts" ^ anInteger << 15 bitAnd: 16rFFFFFFFF! ! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/27/2002 11:25'! temperingShiftU: anInteger "First of four Mersenne Twister shifts" ^ anInteger >> 11 bitAnd: 16rFFFFFFFF! ! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/27/2002 11:13'! xor: aWord1 and: aWord2 "comment stating purpose of message" | t1 mag01 | mag01 _ WordArray with: 0 with: MatrixA. t1 _ (aWord1 bitXor: aWord2 >> 1) bitAnd: 16rFFFFFFFF. ^ (t1 bitXor: (mag01 at: (aWord2 bitAnd: 1) + 1)) bitAnd: 16rFFFFFFFF! ! !MTRandom methodsFor: 'initialization' stamp: 'mwb 7/27/2002 10:26'! initialize "Set up space for seeds, then fill them in" | s | mt _ WordArray new: N. mti _ N+1. [s _ Time millisecondClockValue. self sgenrand: s. s = 0] whileTrue! ! !MTRandom methodsFor: 'accessing' stamp: 'mwb 7/27/2002 09:47'! next "Answer a random Float in the interval [0 to 1)." ^ (self nextValue / 16rFFFFFFFF) asFloat! ! !MTRandom methodsFor: 'accessing' stamp: 'mwb 7/26/2002 17:37'! nextInt: anInteger "Answer a random integer in the interval [1, anInteger]." ^ (self next * anInteger) truncated + 1! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/27/2002 11:16'! genrand | y | self mti >= N ifTrue: [self mti = (N + 1) ifTrue: [self sgenrand: 4357]. 1 to: N - M do: [:kk | y _ self mask: (self mt at: kk) and: (self mt at: kk + 1). self mt at: kk put: (self xor: (self mt at: kk + M) and: y)]. N - M + 1 to: N - 1 do: [:kk | y _ self mask: (self mt at: kk) and: (self mt at: kk + 1). self mt at: kk put: (self xor: (self mt at: kk + (M - N)) and: y)]. y _ self mask: (self mt at: N) and: (self mt at: 1). self mt at: N put: (self xor: (self mt at: M) and: y). mti _ 1]. y _ self mt at: mti. mti _ mti + 1. y _ (y bitXor: (self temperingShiftU: y)) bitAnd: 16rFFFFFFFF. y _ ((y bitXor: (self temperingShiftS: y)) bitAnd: TemperingMaskB) bitAnd: 16rFFFFFFFF. y _ ((y bitXor: (self temperingShiftT: y)) bitAnd: TemperingMaskC) bitAnd: 16rFFFFFFFF. y _ (y bitXor: (self temperingShiftL: y)) bitAnd: 16rFFFFFFFF. ^ y! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/26/2002 11:38'! mt ^ mt! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/26/2002 16:23'! mti ^ mti! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/26/2002 16:23'! mti: anInteger mti _ anInteger! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/26/2002 20:56'! nextValue "This method generates random instances of Integer. The algorithm is described in detail by M. Matsumoto and T. Nishimura in 'Mersenne Twister: A 623-dimensionally equidistributed uniform pseudorandom number generator', ACM Trans. on Modeling and Computer Simulation Vol. 8, No. 1, January pp.3-30 1998." ^ self genrand! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/27/2002 11:25'! seed: anInteger seed _ anInteger. self sgenrand: (anInteger bitAnd: 16rFFFFFFFF)! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/26/2002 17:36'! sgenrand: anInteger mt at: 1 put: (anInteger bitAnd: 16rFFFFFFFF). 2 to: N do: [:i | self mt at: i put: (69069 * (self mt at: i - 1) bitAnd: 16rFFFFFFFF)]. self mti: N ! ! "-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- "! MTRandom class instanceVariableNames: ''! !MTRandom class methodsFor: 'instance creation' stamp: 'mwb 7/26/2002 10:54'! new "Create new instance of MTRandom" ^ super new initialize! ! !MTRandom class methodsFor: 'class initialization' stamp: 'mwb 7/27/2002 11:23'! initialize " MTRandom initialize " "Set period parameters" N _ 624. M _ 397. MatrixA _ 16r9908B0DF. LowerMask _ 16r7FFFFFFF. UpperMask _ 16r80000000. "Set tempering parameters" TemperingMaskB _ 16r9D2C5680. TemperingMaskC _ 16rEFC60000! ! MTRandom initialize! C program: /* A C-program for MT19937: Real number version */ /* genrand() generates one pseudorandom real number (double) */ /* which is uniformly distributed on [0,1]-interval, for each */ /* call. sgenrand(seed) set initial values to the working area */ /* of 624 words. Before genrand(), sgenrand(seed) must be */ /* called once. (seed is any 32-bit integer except for 0). */ /* Integer generator is obtained by modifying two lines. */ /* Coded by Takuji Nishimura, considering the suggestions by */ /* Topher Cooper and Marc Rieffel in July-Aug. 1997. */ /* This library is free software; you can redistribute it and/or */ /* modify it under the terms of the GNU Library General Public */ /* License as published by the Free Software Foundation; either */ /* version 2 of the License, or (at your option) any later */ /* version. */ /* This library is distributed in the hope that it will be useful, */ /* but WITHOUT ANY WARRANTY; without even the implied warranty of */ /* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. */ /* See the GNU Library General Public License for more details. */ /* You should have received a copy of the GNU Library General */ /* Public License along with this library; if not, write to the */ /* Free Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA */ /* 02111-1307 USA */ /* Copyright (C) 1997 Makoto Matsumoto and Takuji Nishimura. */ /* Any feedback is very welcome. For any question, comments, */ /* see http://www.math.keio.ac.jp/matumoto/emt.html or email */ /* matumoto(a)math.keio.ac.jp */ #include<stdio.h> /* Period parameters */ #define N 624 #define M 397 #define MATRIX_A 0x9908b0df /* constant vector a */ #define UPPER_MASK 0x80000000 /* most significant w-r bits */ #define LOWER_MASK 0x7fffffff /* least significant r bits */ /* Tempering parameters */ #define TEMPERING_MASK_B 0x9d2c5680 #define TEMPERING_MASK_C 0xefc60000 #define TEMPERING_SHIFT_U(y) (y >> 11) #define TEMPERING_SHIFT_S(y) (y << 7) #define TEMPERING_SHIFT_T(y) (y << 15) #define TEMPERING_SHIFT_L(y) (y >> 18) static unsigned long mt[N]; /* the array for the state vector */ static int mti=N+1; /* mti==N+1 means mt[N] is not initialized */ /* initializing the array with a NONZERO seed */ void sgenrand(seed) unsigned long seed; { /* setting initial seeds to mt[N] using */ /* the generator Line 25 of Table 1 in */ /* [KNUTH 1981, The Art of Computer Programming */ /* Vol. 2 (2nd Ed.), pp102] */ mt[0]= seed & 0xffffffff; for (mti=1; mti<N; mti++) mt[mti] = (69069 * mt[mti-1]) & 0xffffffff; } /* double */ /* generating reals */ unsigned long /* for integer generation */ genrand() { unsigned long y; static unsigned long mag01[2]={0x0, MATRIX_A}; /* mag01[x] = x * MATRIX_A for x=0,1 */ if (mti >= N) { /* generate N words at one time */ int kk; if (mti == N+1) /* if sgenrand() has not been called, */ sgenrand(4357); /* a default initial seed is used */ for (kk=0;kk<N-M;kk++) { y = (mt[kk]&UPPER_MASK)|(mt[kk+1]&LOWER_MASK); mt[kk] = mt[kk+M] ^ (y >> 1) ^ mag01[y & 0x1]; } for (;kk<N-1;kk++) { y = (mt[kk]&UPPER_MASK)|(mt[kk+1]&LOWER_MASK); mt[kk] = mt[kk+(M-N)] ^ (y >> 1) ^ mag01[y & 0x1]; } y = (mt[N-1]&UPPER_MASK)|(mt[0]&LOWER_MASK); mt[N-1] = mt[M-1] ^ (y >> 1) ^ mag01[y & 0x1]; mti = 0; } y = mt[mti++]; y ^= TEMPERING_SHIFT_U(y); y ^= TEMPERING_SHIFT_S(y) & TEMPERING_MASK_B; y ^= TEMPERING_SHIFT_T(y) & TEMPERING_MASK_C; y ^= TEMPERING_SHIFT_L(y); /* return ( (double) y / (unsigned long)0xffffffff ); */ /* reals */ return y; /* for integer generation */ } /* this main() outputs first 1000 generated numbers */ main(int argc, char* argv[]) { int j; sgenrand(4357); /* any nonzero integer can be used as a seed */ for (j=0; j<1000; j++) { printf("%5lu\n", genrand()); if (j%8==7) printf("\n"); } printf("\n"); }

2 1

[ANN] [UPDATES] 20 updates for 3.3a
by Scott Wallace 28 Jul '02

28 Jul '02

Here are the preambles for twenty updates for Squeak 3.3a that have recently been released to the external 3.3a update stream. Cheers, -- Scott ---------------------- 4915resourceColorFix-raa -- Bob Arning -- 23 June 2002 Republished 7/7/02 sw to the 3.3a update stream, with preamble updated. When loading resource Forms, do NOT discard color table information. It was only with the advent of update 4859 (3.3a) that this became an issue." 4916dupKeys-nk -- Ned Konz -- 11 July 2002 A persuasive argument was made by David Farber to allow duplication of Ctrl and Alt keys. This adds a preference called 'duplicateControlAndAltKeys' that does just that. You can still just swap the keys if you want. Per Doug Way's suggestion, this also changes the default keys to swap to be: $c . $x . $v . $a . $s . $f . $g . $z x (cut) c (copy) v (paste) a (select all) s (save) f (find) g (find again) z (undo) And (unlike the original change set) this will maintain existing Preference settings when installed. The two preferences are mutually exclusive; setting one will reset the other. One small manual correction made in the Postscript by Scott Wallace 7/11/02. 4917dupNavBar-sw -- Scott Wallace -- 11 July 2002 Removes the old navigator bar on reloaded projects if the current image requests suppression of the navbar." 4918loadEToys-sw -- Scott Wallace -- 11 July 2002 Removes barriers to loading etoys into 3.3. Caution: this update is for 3.3 only." 4919reportFullStack-sw -- Scott Wallace -- 19 July 2002 In the error report, print out the short form of *all* stack frames in the same 'rcvr(class) >> selector' format, after printing out the long form of the first four stack frames. This makes the entire compact summary available in a single, readable tableau. Formerly, the first four stack frames did not appear in the compact summary." 4920flapTuning-sw -- Scott Wallace -- 19 July 2002 * Gives explicit object-names to the etoy flaps (Supplies & Navigator) so that they show up in the morph halo that way, per request from Kim Rose. * Makes the Navigator shrink-wrapped, as per prescription from Andreas." 4921scriptParameters-sw -- Scott Wallace -- 20 July 2002 Allows any user-defined etoy script to have a parameter. To add a parameter, invoke 'add parameter' from Scriptor or Viewer menu. Change parameter *type* from the Scriptor header, using up/down arrows or clicking for menu. Obtain a parameter tile by dragging from the parameter position in the Scriptor header. Caution: This applies to 'classic tiles' only. 'Universal tiles' are not handled by this work. Note: the following is a hand-jimmied class-definition for ParameterTile to make it ok both on 3.2 and 3.3" 4922paramFixups-sw -- Scott Wallace -- 22 July 2002 * Fixes a glitch in backing out of renaming a script -- we had not been allowing the user to retain an existing name if it had a colon in it. * If you change a parameter type in such a manner that it introduces a type error for any tile phrases in the script that called for the parameter when it was a different type, the tiles fix themselves up to keep everything type-rosy. * Adds a menu item to TilePad's halo menu allowing you to restore the default literal tile for the particular type. This allows you, albeit somewhat obscurely, to revert a Parameter tile to a plain numeric tile, for example." 4923optionalLexiconPane-sw -- Scott Wallace -- 23 July 2002 Makes the custom Lexicon control buttons occupy a separate pane *in addition to* the one governed by the optionalButtons preference. Thus, if #optionalButtons is true, a Lexicon will have *two* rows of control buttons -- one offering the custom lexicon-specific navigation features and the other with the standard repertory of senders, implementors, etc, which had for space reasons previously had to be omitted from the Lexicon's touchable controls. Also, fixes a seemingly-long-standing bug that appears to have made protocol browsers not work properly in mvc." 4924findA-sw -- Scott Wallace -- 23 July 2002 Consolidates the main work of six methods that service the various windows-menu 'find-a' requests (e.g. 'find a Transcript') into a single place. One of these six is the newly added 'find a fileList', with cmd-L as its desktop-cmd-key shortcut." 4925fixAllCat-sw -- Scott Wallace -- 24 July 2002 Fix up methods that had '-- all --' as their home category" 4926versionBrowser-BG -- Boris Gaertner, Scott Wallace -- 25 July 2002 Fixes a glitch that caused the creation of a spurious concrete method category named '-- all --' when a method definition was reverted in a versions browser that had been originally launched from a browser that had had the '-- all --' category selected. Note: Boris penetrated to the core of this bug, and then produced a fix on a Squeak 2.7 system; Scott integrated Boris's fix with the post-2.7 changes to the method, and did some further trimming." 4927condenseSources-sw -- Scott Wallace -- 25 July 2002 Because the presence of spurious concrete '-- all -- method categories breaks the condenseSources process, we now make a preemptive call to the method that fixes up all such errors before proceeding with the real work of condenseSources. Thanks to Boris Gaertner for finding this problem." 4928chgSetForVersion-dew -- Doug Way -- 7 September 2001 Adds a 'find original change set' menu item for the Versions Browser, which locates the single changeset (or sources file) which originally contained the selected method version. It does this by matching the file pointers stored in the ChangeRecord and ChangeSet." 4929parameterMenuIcon-sw -- Scott Wallace -- 27 July 2002 Use a black triangle as the type-menu icon on a parameter tile, as per suggestion from Alan." 4930fasterSlideToTrash-sw -- Scott Wallace -- 26 July 2002 When sliding large morphs to the trash, animate a thumbnail rather than the full-sized morph. This speeds things up considerably." 4931versBrowsFix-sw -- Scott Wallace -- 27 July 2002 Fixes a glitch in update 4926 that only showed up when browsing versions of a method on the class side." 4932changeEnh-nk -- Ned Konz -- 26 June 2002 This adds an item to the ChangeSorter category menu that lets you make a change-set category with all the change sets that change a particular class." 4933changedOthersMethods-nk -- Ned Konz <ned(a)bike-nomad.com> -- 30 March 2002 This adds a choice to the shifted change set menu in ChangeSorters and ChangeSetBrowsers (called 'check for any other authors') that will find any methods with historical versions that don't have your initials. This can be used to find other people's code that you've modified." 4934chgSorterMisc-sw -- Scott Wallace -- 27 July 2002 * Adds check-for-uncommented-classes to the checks section of the shifted change-sorter menu. * Adds cmd-hypen as kbd shortcut for subtract-other-side in the dual change sorter. * Adds ClassListBrowser, a browser variant that shows the code for an arbitrary list of classes. See class comment or browse class-side code for examples"

1 0

← Newer
1
...
5
6
7
8
9
10
11
...
40
Older →

Jump to page: